Video GenerationActive
blackforestlabs / flux-3-v2v
flux-3-v2v
Extend a source video past its final frame. FLUX 3 V2V generates 5–20 seconds of 720p or 1080p footage, with optional synced audio.
Video to Video
Model ID
flux-3-v2v
Provider
blackforestlabs
Updated
1785932815
wiro playground—blackforestlabs/flux-3-v2v
Updated 1785932815
## Overview
FLUX 3 V2V is a video-to-video continuation model from Black Forest Labs. You upload a short clip, then describe what happens next. The model predicts new frames that carry the scene forward from the last frames of your source video. It can also generate a matched soundtrack in the same pass, so audio events line up with on-screen motion. This helps you add usable extra seconds to a shot without manual editing or frame-by-frame work.
## What you can build
- Extend an establishing shot to fit a longer edit without a hard cut
- Add an outro beat to a product clip, like a logo reveal or pack shot
- Continue a camera move past the end of the source footage
- Create extra b-roll seconds from an existing clip for social variants
- Make smoother transitions by extending the first clip before the cut
- Generate alternate endings from the same input clip and story direction
## Inputs
- A source video clip, provided as an upload or a public URL. Use MP4 for best compatibility. Keep it under 50 MB and under 15 seconds.
- A required text description of what should happen after the uploaded video ends. Mention actions, camera motion, and key sound cues.
- An optional choice of output duration. You can let the model choose, or set a whole number of seconds from 5 to 20.
- An optional output resolution choice. Pick 720p for standard HD output or 1080p for full HD.
- An optional aspect ratio setting. You can let the model choose, or force 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, or 9:16.
- An optional audio toggle. Leave it on for a soundtrack, or turn it off for silent video.
- An optional safety strictness level from 0 to 4. Lower values apply stricter filtering.
## Outputs
- A generated MP4 video file you can download from a hosted URL.
- The video can include an audio track when audio generation is enabled.
- A numeric seed value that identifies the exact generation. Save it for repeatability in your workflow.
## Limitations
- Input clips must stay under 15 seconds and under 50 MB.
- Output duration is limited to short clips. Use 5 to 20 seconds.
- Output resolution is limited to 720p or 1080p.
- Aspect ratio choices are limited to the listed set.
- Continuations can drift if the input has fast cuts, heavy motion blur, or abrupt lighting changes.
- Low-quality inputs increase failure risk. This includes shaky footage, compression artifacts, and very dark scenes.
- If your prompt contradicts the last frames, the model may change details or introduce new objects.
## Safety & compliance
- Don’t use this model to generate illegal content or non-consensual intimate imagery.
- Child sexual abuse material is prohibited. Platforms commonly apply detection and reporting safeguards.
- Requests may be filtered by automated moderation across text and video.
- Higher safety strictness can reduce risky outputs but may also block borderline prompts.
- You’re responsible for rights and permissions for the source video you upload.
Example prompts
Great starting points for flux-3-v2v.
Cinematic forward tracking shot continuing to follow the subject. The monk in the bright orange robe continues his slow, peaceful walk further down the dirt path, gradually moving deeper into the garden. A light breeze subtly rustles the foliage of the dark trees on the left and the plants draped over the stone wall on the right. Soft, natural overcast lighting remains steady. Photorealistic.Video Generation
Cinematic static shot. The dark brown horse continues to graze peacefully, its head lowered into the bright yellow flowers. It takes a slow, heavy step forward through the thick green grass, subtly swishing its dark tail. A gentle breeze ripples across the vibrant meadow, causing the flowers and the dense green trees in the background to sway softly in the bright natural sunlight. Photorealistic, nature documentary style.Video Generation
Cinematic static shot. The tail end of the red and white commuter train completely exits the frame to the right, leaving the railway tracks empty. The massive, weathered steel arched bridges stand silently under the clear, bright blue sky. A gentle breeze slightly rustles the dry, brown grass and bare bushes in the foreground along the green chain-link fence. Photorealistic, urban documentary style.Video Generation
Cinematic close-up, slow push-in. The ladle finishes pouring the clear liquid into the hot pan, instantly triggering a thick eruption of white steam. The chef gently shakes the pan back and forth over the flickering blue gas flame. Moody kitchen lighting highlights the rising vapor. Photorealistic.Video Generation
API quick start
Run flux-3-v2v with a single API call.
POST https://api.wiro.ai/v1/Run/blackforestlabs/flux-3-v2v
{
"prompt": "Cinematic static shot. The dark brown hor…",
"inputVideo": "https://your-cdn.com/input.mp4",
"duration": 5,
"resolution": "hd"
}